Varying dietary EPA/DHA ratios influence growth and viral disease response in Atlantic salmon (Salmo salar L.) fry

Abstract

Balancing dietary n−3 long-chain polyunsaturated fatty acids (LC-PUFAs) is essential for optimizing growth and disease resilience in Atlantic salmon and will become an increasing challenge when implementing novel sources of eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), such as genetically modified (GM) oils and microalgae oils. To study this, Atlantic salmon (Salmo salar L.) fry (initial weight 0.15 g) were pre-fed diets differing in EPA/DHA ratios (0.2 E/D, 0.9 E/D, and 2.2 E/D) while maintaining constant total EPA+DHA for 4 weeks. Following this, and while kept on the same dietary treatments, the fry were challenged with wild-type or attenuated salmonid alphavirus (SAV-WT and SAV-Att, respectively). The duration of the disease challenge was up to 28 days, with a final fish weight of approximately 1.1 g at the end of the trial. There was a clear decrease in fatty acid storage during SAV-WT infection, which was not observed during SAV-Att infection. We demonstrate a preferential decrease in monounsaturated fatty acids during SAV-WT infection, preserving the n−3 LC-PUFA content in the tissues. Increasing the dietary EPA/DHA ratio significantly reduced growth performance, indicating that the dietary balance of EPA to DHA influences growth during the start feeding of salmon fry. Following infection, dietary effects were outcome-dependent. Fish fed the low EPA/DHA diet (0.2 E/D) showed overall higher viral loads, but improved histopathology at the end of the trial. In contrast, fish fed the high EPA/DHA diet (2.2 E/D) exhibited signs of improved histopathology at the end of the trial, yet experienced the highest (non-significant) mortality after SAV-WT challenge. Fish fed the balanced diet (0.9 E/D) showed poorer histopathological scores at the end of the trial, but did not display elevated mortality. These findings demonstrate that dietary EPA/DHA ratios differentially modulate growth and disease outcome, exerting distinct and sometimes opposing effects on viral load, tissue inflammation, and survival. The results highlight the complexity of optimizing n−3 LC-PUFA balance to support both performance and antiviral resilience in Atlantic salmon.

1 Introduction

Fish oil (FO) and fishmeal (FM) have traditionally provided farmed Atlantic salmon (Salmo salar L.) with the essential long-chain omega−3 fatty acids (FAs) eicosapentaenoic acid (EPA; 20:5n−3) and docosahexaenoic acid (DHA; 22:6n−3) in near equal proportions (Tocher et al., 2019; Glencross et al., 2024). A growing global demand for FO has led to increased replacement with plant oils over the past two decades (Aas et al., 2022; Glencross et al., 2024). However, further reducing FO inclusion from current levels likely includes the risk of welfare costs (Glencross et al., 2025), which, combined with concerns of the sustainability and access to marine feed resources, has encouraged the development of other dietary sources of EPA and DHA (Tocher et al., 2019; Oliver et al., 2020; Liu et al., 2022). Many alternative EPA and DHA sources exhibit variable EPA/DHA ratios compared to the conventional FO (Tocher et al., 2019). Such variability is noteworthy, as recent studies have demonstrated that the dietary EPA/DHA ratio may affect growth performance, oxidative stress response, and immune function across multiple fish species (Glencross et al., 2014; Betancor et al., 2016; Bou et al., 2017; Santigosa et al., 2023; Hundal et al., 2026). However, other studies have suggested no effect on performance with differing dietary EPA/DHA ratios (Glencross et al., 2014; Betancor et al., 2016; Bou et al., 2017; Santigosa et al., 2023). The effects on the growth of EPA/DHA are therefore hypothesized to be context-dependent and closely linked to how the fish overall respond to different stressors, thus necessitating a physiological challenge to fully showcase the effects of varying this parameter.

EPA and DHA perform different functions in the body. DHA is important structurally, particularly in neural and retinal tissues (Sissener et al., 2016), and is preferentially incorporated into and retained in cellular membranes rather than oxidized for energy or converted into other FAs (Turchini et al., 2009; Tocher, 2010). DHA is also an important precursor for specialized pro-resolving mediators (SPMs) (Ferreira et al., 2022) and is thus important in inflammation resolution. Meanwhile, EPA is not very efficiently stored and is more often used for β-oxidation than DHA (Norambuena et al., 2015). It is also used for the endogenous production of DHA through elongation and desaturation (Oboh et al., 2017) or metabolized into eicosanoids and SPMs (including E-series resolvins) (Serhan and Petasis, 2011; Calder, 2017). While both EPA and DHA therefore contribute to the resolution of inflammation, they do so via distinct mediator families and pathways, with differences in functions and potencies (Basil and Levy, 2016; Araujo et al., 2025). Changes in the dietary EPA/DHA ratio can also alter the balance between n−6 and n−3 FAs in polar lipids of Atlantic salmon tissues (Hundal et al., 2026), which is also shown to affect the inflammatory response in salmon (Araujo et al., 2014; Hundal et al., 2021). Alterations in the dietary FA profile, particularly in the levels of and ratios between EPA, DHA, and arachidonic acid (ARA), can therefore influence the balance between pro- and anti-inflammatory lipid mediators, potentially modulating the viral immune response as suggested in earlier studies (Martinez-Rubio et al., 2012; Caballero-Solares et al., 2017).

Salmonid alphavirus subtype 3 (SAV3) causes pancreas disease (PD) in farmed Atlantic salmon and rainbow trout and shares ~91%–93% nucleotide identity with other SAV subtypes (Taksdal et al., 2007; Fringuelli et al., 2008). Infected fish typically show pancreatic necrosis and myocarditis, and PD ranks among the leading causes of seawater mortality in Norwegian aquaculture (Taksdal et al., 2007; Norwegian Veterinary Institute, 2023). In experimental settings, wild-type SAV3 (SAV-WT) has been shown to cause acute PD with severe pathology, including high viral loads and extensive tissue damage (Braaen et al., 2025). We hypothesize that such strong virulence can mask more subtle effects of dietary treatments in some experimental setups due to practical constraints on sample numbers and because physiological improvements may not be observable before death. In contrast to SAV-WT, attenuated SAV3 (SAV-Att) strains engineered for reduced virulence while maintaining tissue tropism have demonstrated milder clinical outcomes and no or much reduced mortality (Aksnes et al., 2021; Braaen et al., 2025). These models possibly offer a more sensitive system for studying the effects of feed composition on host immune responses, enabling the detection of nuanced physiological and immunological changes that may be masked by the lethality of SAV-WT. The lower virulence of attenuated viruses tips the balance in favor of the host, likely allowing more controlled and optimal host responses. Accordingly, attenuated grass carp reovirus (GCRV) has been shown to cause less disruption to host metabolism than virulent wild-type GCRV in experimental challenges (Ma et al., 2023). Salmon fry provide a useful, relatively novel model for studying infections in salmon. They are at least as sensitive to SAV3 infection as older stages (Kannimuthu et al., 2026) and can distinguish between differently attenuated variants, suggesting distinct infection kinetics and outcomes in terms of growth, morbidity, and mortality (Kannimuthu et al., 2026; S. Grove, pers. comm.), although such observations remain to be formally validated.

Several studies have investigated the effect of EPA on inflammation and immunity in Atlantic salmon. Gjøen et al. (2024) pre-incubated Atlantic salmon head kidney (ASK) cells with increasing levels of EPA to investigate its effects on immune and anti-viral responses following stimulation with viral mimic polyinosinic:polycytidylic acid (poly I:C). While EPA had minimal effects on immune gene expression in unstimulated cells, significant effects of EPA were found in poly I:C-challenged cells, including indications of reduced inflammatory processes (Gjøen et al., 2024). Similarly, in ASK cells pre-incubated with EPA and inoculated with infectious salmon anemia virus (ISAV), EPA had a limited effect on the innate immune mechanisms alone, but during infection, many processes were affected (Holmlund et al., 2025). Furthermore, the EPA has demonstrated beneficial effects against other viral infections in salmon. For example, in a study by Hatlen et al. (2016), fish fed diets high in EPA experienced no mortality during an outbreak of heart and skeletal muscle inflammation [HSMI; causative agent is piscine orthoreovirus (PRV)], in contrast to the significant mortality detected in other dietary groups. They observed similar trends of improvements related to dietary EPA levels for PD (Hatlen et al., 2016). Similarly, trials with functional feeds high in EPA have shown promising results in reducing inflammation and heart lesions associated with cardiomyopathy syndrome (CMS) (Martinez-Rubio et al., 2014). However, as these functional feeds involved multiple compositional changes, the specific contribution of EPA remains unclear. Studies performed to research the role of DHA in Atlantic salmon also point to its function in protecting salmon cells against oxidative stress and the production of anti-inflammatory lipid mediators (Holen et al., 2015; Bou et al., 2020; Araujo et al., 2025). However, a feeding trial performed with diets predominant in DHA and low in EPA did not show the same findings as high EPA diets, rather pointing toward positive effects on markers of fish welfare of a more balanced EPA/DHA ratio at 1:1 compared to lower EPA/DHA ratios (Santigosa et al., 2023). These findings collectively point to a protective role of EPA in mitigating disease-induced inflammation. The hypothesis for this study was therefore that a high EPA/DHA ratio would positively impact immune function and disease resilience.

To address these knowledge gaps, the present study investigated how Atlantic salmon fry respond to infection with SAV-WT and SAV-Att when fed diets containing equal levels of total EPA+DHA but varying EPA/DHA ratios. The goal was to determine whether the EPA/DHA ratio influences growth, immune function, and disease resilience, while additionally exploring SAV-Att as a tool in studying nutrient–disease interactions in comparison to SAV-WT.

2 Materials and methods

2.1 Virus preparation

Virus for the experimental challenge was prepared using Chum salmon heart cells (CHH-1) transfected with infectious clones containing either a SAV-WT genome or a recombinant SAV3 genome containing five codon pair bias de-optimized (CPBD) regions (SPDV3₁₂₃₅₆/SAV-Att) (unpublished, pers. comm., Søren Grove). The SAV-WT was constructed by inserting a synthesized genome from a wild-type virus isolate (GenBank KC122926.1) described by Petterson et al. (2013) into a pVAX1 plasmid vector. The five CPBD regions in SAV-Att were designed using the program CodonShuffle (Jorge et al., 2015) followed by manual curation. This approach introduced a total of 882 synonymous nucleotide changes compared to the wild-type genome, ensuring that the codon pair bias was distinctly reduced in the recoded regions. Importantly, the recoding kept genomic characteristics, including the effective number of codons, codon adaptation index, the free energy of RNA folding, and the frequency of CpG and UpA dinucleotides, as close as possible to the original values in the wild-type genome (pers. comm., Søren Grove). Compared to SAV-WT, earlier studies have shown that the SAV-Att significantly reduced viral replication in vitro and led to no mortality and lower viral loads than SAV-WT in challenged Atlantic salmon fry (pers. comm., Søren Grove).

2.2 Diet formulation and production

Three experimental diets were produced by Cargill Innovation Center (CIC, Dirdal, Norway) and were formulated to be iso-nitrogenous (~54% crude protein) and iso-lipidic (~16% crude lipid). All diets were based on an identical basal pellet formulation, differing only in the type and amount of EPA oil included. All ingredients, including the oils and premix, were included in the basal pellet mixture. The diets were extruded using a Wenger X85 single-screw extruder, custom-designed in 1996 for pilot-scale feed production. The extrusion was performed at 100°C for a minimum of 30 s and dried (30 min at 90°C for the first 10 min and then decreased gradually to 70°C, where it was kept for the last 10 min). The pellet was further crushed and sieved to crumbles of ~0.8–1 mm on a mobile sieving machine. The three diets were designed to have equal EPA+DHA, but with varying EPA/DHA (E/D) ratios. The first diet (0.2 E/D) contained a low relative level of EPA, the second (0.9 E/D) had a balanced EPA/DHA ratio, and the third (2.2 E/D) had a high relative level of EPA. Additionally, the diets were formulated with a similar sum of saturated FAs (SFAs), monounsaturated FAs (MUFAs), and polyunsaturated FAs (PUFAs). Diet formulation and the analyzed proximate composition can be found in Table 1.

2.3 Fish trial

Approximately 3,000 Atlantic salmon fry were procured from Lerøy Sjøtroll Kjærelva AS (Fitjar, Norway) immediately after the internalization of the yolk sac. The fish were transferred to the Institute of Marine Research’s (IMR’s) disease challenge facility (Bergen, Norway) and stocked in three tanks (1,000 fish tank⁻¹), containing 400-L aerated fresh water at 13°C. The fish were maintained on a 12:12-h light:dark regime. From the onset of first feeding, the salmon fry groups were fed one of three experimental diets ad libitum three times a day for a period of 4 weeks (pre-feeding, overview of experimental setup in Figure 1). After 4 weeks, the three dietary groups of salmon fry were anesthetized in phosphate buffer with 50 µg L⁻¹ of tricaine methanesulfonate (Finquel, MSD Animal Health Norway, Bergen, Vestland) and challenged by intraperitoneal (IP) injection of 20 µL of either SAV-WT or SAV-Att virus suspension of titer 10^3.6 TCID₅₀ mL⁻¹ and divided into 18 tanks (8 L, 60 fish tank⁻¹). Non-infected controls were injected with 20 µL of L-15 cell culture medium from non-infected CHH-1 cells and divided into nine tanks (8 L, 55 fish tank⁻¹). One day post-challenge (1 DPC), ad libitum feeding with experimental diets was resumed for the remainder of the experiment. All the experimental procedures were reviewed and approved by the Norwegian Animal Research Authority prior to the start of challenge experiments (FOTS ID:30463).

Figure 1

2.4 Samplings

Salmon fry were sampled at multiple time points before and during the experimental trial (see Figure 1; Supplementary Table 1 for more details on sample sizes per sample type and sampling/treatment). Prior to each sampling, fish were euthanized in phosphate buffer with 200 µg L⁻¹ of tricaine methanesulfonate (Finquel, MSD Animal Health). Upon arrival at the IMR’s disease challenge facility, the fish were weighed, but due to the small size of the fry, 30 fish were counted and weighed together. This was done three times with different fish each time, giving an estimated initial weight of the fish at 0.15 ± 0.00 g. Further, 30 fish were pooled into one sample for the analysis of FA composition. After 4 weeks of pre-feeding with experimental diets and prior to challenge (i.e., 0 DPC), fish were sampled for the analysis of FA composition (pooled samples), viral load (individual samples), and histopathology (individual samples). Samplings at 7 and 14 DPC were performed for all treatment and dietary groups, while the final sampling differed between the SAV-WT and SAV-Att groups (Figure 1). Considering the rapid progress of disease in SAV-WT, these fish were sampled at 19 DPC due to already large mortalities occurring in the days before that time point. Non-infected controls were also sampled the following day (20 DPC) to have comparable control samples. The SAV-Att and their non-infected controls were sampled at 28 DPC, which was the planned maximum duration of the disease challenge. At each time point, samples were collected for the analyses of FA composition (pooled samples), viral load (individual samples), and histology (individual samples). Consequently, the sampling protocol was designed to increase the number of sampled fish at later samplings to compensate for the expected increase in individual variability. An overview of the experimental design and sampling is given in Figure 1, and specific details on samples collected for each time point and analyses are summarized in Supplementary Table 4.

2.5 Chemical composition of the diets

The chemical composition of the diets was analyzed following Association of Official Agricultural Chemists (AOAC) (2010) methods. Samples were freeze-dried for 48 h (FreeZone 18 Liter Console, Labconco, Kansas City, MO, USA) to obtain dry matter and then ground into a fine powder for the analysis of nitrogen (N) and fat content. Nitrogen content was measured using a CHNS elemental analyzer (Vario Macro Cube, Elementar Analysensysteme GmbH, Langenselbold, Germany) and quantified according to AOAC (2010). The analyzer was calibrated with ethylenediaminetetraacetic acid (Leco Corporation, Saint Joseph, MI, USA), while sulfanilamide (Alfa Aesar GmbH & Co, Karlsruhe, Germany) and a standard meat reference material (SMRD 2000, LGC Standards, Teddington, UK) served as controls. Crude protein was calculated as N × 6.25. Crude fat content (via ethyl acetate extraction) was measured gravimetrically.

2.6 FA composition

The FA composition of the diets and homogenized whole salmon fry was analyzed following the method of Lie and Lambertsen (1991) via gas–liquid chromatography (Scion 436-GC, Scion Instruments, Livingston, Scotland, UK). Lipids were extracted from the samples using a chloroform/methanol mixture (2:1, v/v), followed by filtration. The filtered lipid extracts were then saponified and methylated using 12% BF₃ in methanol. FAs were identified based on retention times using a standard mixture of methyl esters (Nu-Chek-Prep, Elysian, MN, USA) and quantified with 19:0 as the internal standard. The Chromeleon^® software version 7.2 (Thermo Scientific, Waltham, MA, USA) was used to integrate chromatographic peak areas.

2.7 RNA extraction and RT-qPCR

Total RNA was extracted from organ packages using the Promega ReliaPrep simplyRNA HT 384 kit on a Biomek 4000 Automated Workstation following the manufacturer’s protocol. RNA concentration was measured using a NanoDrop™ 8000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA), and samples were normalized to 50 ng/μL using the Biomek 4000. The total RNA was then used in a quantitative RT-qPCR assay targeting the SAV3 nsp1 gene, with AgPath-ID One Step RT-PCR reagents (Hodneland and Endresen, 2006) (Supplementary Table 4). Each 10-μL reaction, run on a 384-well plate, contained 100-ng RNA, 900-nM forward and reverse primers, and 150-nM probe. The Atlantic salmon elongation factor α 1 gene (EF1α) was used as an internal control (Olsvik et al., 2005). Amplification and fluorescence detection were carried out on a QuantStudio™ 5 Real-Time PCR System for 40 cycles. Samples with cycle threshold (Ct) values of 40 were considered negative in figures and statistical analyses.

2.8 Histopathology

Whole fry (tails removed) were fixed in 4% formaldehyde solution for 48 h and transferred to 70% ethanol. After being dehydrated using a graded ethanol series and cleared by xylene in an Automatic Benchtop Tissue Processor (Leica Biosystems, Nußloch, Germany), they were embedded in paraffin and sectioned at a thickness of 3 μm. Sections were stained with a combination of Shandon instant hematoxylin (Thermo Scientific), erythrosine, and saffron (HES stain; Waldeck, Münster, Germany). Scanning was performed using a NanoZoomer S60 digital slide scanner (Hamamatsu Photonics, UK), and images were viewed in the NanoZoomer Digital Pathology NDP.view2 software (Hamamatsu Photonics, Hertfordshire, UK). Selected sections were imaged using a Leica DMRBE microscope fitted with a SPOT Insight camera. The scoring system established by McLoughlin et al. (2006) was employed to evaluate the histopathology (Supplementary Table 4).

2.9 Statistical analysis

Statistical analyses were performed using the free software environment R (v4.3.3) (R Core Team, 2024). Growth was analyzed using linear models with diet as a fixed effect at 0 DPC, and differences were assessed using one-way ANOVA. Since these samples were taken from a single pre-feeding tank per diet, there was no tank effect at this time point. For results at 20 and 28 DPC, diet and infection status (control/infected) were used as fixed effects with tank as a random effect and assessed using two-way ANOVA. Post-hoc pairwise comparisons with Tukey’s adjustment were performed using the emmeans package (Lenth, 2025). Linear models were fitted for FA composition with diet and infection status as fixed effects. Since these samples were pooled per tank to have enough sample material for analyses, there was also no tank effect taken into account for this parameter. Model residuals met assumptions of normality and homoscedasticity, tested using the Shapiro–Wilk test (base R) and Levene’s test performed with the car package by Fox and Weisberg (2019). Fixed effects and their interaction were assessed using two-way ANOVA. Significant main effects were followed up by post-hoc pairwise comparisons with Tukey’s adjustment using the emmeans package (Lenth, 2025). Viral load was analyzed using a linear mixed model with diet, DPC, and virus type as fixed effects and tank as a random effect [lme4 (Bates et al., 2015) and lmerTest packages (Kuznetsova et al., 2017)]. Model complexity was determined using likelihood ratio tests, retaining the three-way interaction model. Fixed effects were tested using Type III Wald chi-square tests, and post-hoc pairwise comparisons used Tukey’s adjustment with the emmeans package (Lenth, 2025). Standard residual diagnostics indicated moderate non-normality (Shapiro–Wilk W = 0.928) and mild heteroscedasticity, assessed using QQ plots and residuals vs. fitted plots, respectively. Bootstrap confidence intervals (n = 10,000) confirmed the three-way interaction structure to be robust to these deviations. Fifty SAV-Att observations (13.6%) were at the Ct detection limit, distributed across treatment groups, and retained in the analysis. Histopathological scores were analyzed using cumulative link mixed models (CLMM) with a logit link function using the ordinal package (Christensen, 2025). Diet and DPC were set as fixed effects and the tank as a random effect. The proportional odds assumption was assessed using nominal and scale tests. Where minor violations were detected or could not be fully evaluated due to the restricted score distributions, the CLMM was retained, as it is robust to moderate violations (Harrell, 2015).

Statistical significance was set at p < 0.05. Figures and graphs for viral load and histology were made using GraphPad Prism 10 (v10.3.1) (GraphPad Software, La Jolla, CA, USA; www.graphpad.com). FA figures were created in R using packages tidyverse (Wickham et al., 2019), ggplot2 (Wickham, 2016), and patchwork (Pedersen, 2025).

3 Results

3.1 Diet composition

All diets had a similar total level of EPA and DHA [~29% of total FA (TFA)], differing only in their EPA/DHA ratios. The ratios were 0.2, 0.9, and 2.2, and the diets will be referred to as 0.2 E/D, 0.9 E/D, and 2.2 E/D, respectively. The 0.9 E/D and 2.2 E/D diets contained approximately three and five times more EPA than the 0.2 E/D diet (4.4%, 13.7%, and 20.2% of TFA in the 0.2 E/D, 0.9 E/D, and 2.2 E/D diets, respectively); consequently, DHA levels were 24.1%, 15.9%, and 9.0% of TFA, respectively. As formulated, the three experimental diets had the same ΣSFA (~10.5% of TFA), ΣMUFA (~35.5% of TFA), ΣPUFA (~49% of TFA), and ARA/EPA (~0.07). A slight decrease in dietary n−3/n−6 was seen with the increase in EPA/DHA. The concentration of docosapentaenoic acid (DPA; 22:5n−3) also varied among the experimental diets, decreasing with increasing dietary EPA levels (DPA at 3.5%, 2.1%, and 1.0% of TFA in the 0.2 E/D, 0.9 E/D, and 2.2 E/D diets, respectively). The FA composition of experimental diets is given in Table 2.

3.2 Growth and performance

Feed intake was not measured accurately due to the limitations of the rearing system and is therefore not reported. After 4 weeks of pre-feeding (i.e., at 0 DPC, before IP challenge), salmon fry fed the 2.2 E/D diet had a significantly lower weight compared to those fed the 0.2 E/D diet, with fish fed the 0.9 E/D diet being intermediate (Table 3). Notably, the non-infected control fish sampled at 20 and 28 DPC displayed similar growth differences as seen after the pre-feeding, although they were not significant.

At 20 DPC, only the SAV-WT groups and their controls were sampled. The fish infected with SAV-WT had a significantly lower weight compared to the non-infected control fish across all diets. Within the control group, fish fed the 0.2 E/D diet showed significantly higher weight compared to those fed the 2.2 E/D diet, whereas no significant differences in body weight were observed among diets within the infected group. No interaction between infection and diet was detected.

At 28 DPC, only the SAV-Att groups and their controls were sampled (Table 3). There was a significant main effect of infection and a significant interaction. Within the 0.2 E/D diet, fish infected with SAV-Att displayed significantly lower weight than non-infected control fish.

As expected, total mortality was high in salmon fry challenged with SAV-WT, with the first mortality recorded at 17 DPC. Although not significant, the mortality in the 2.2 E/D fed fish was nearly double that of the fish fed the 0.2 E/D and 0.9 E/D diets (Table 3). No mortality occurred in the control fish or in fish challenged with SAV-Att (Table 3).

3.3 FA composition

Fatty acid results are presented as both relative proportions (% of total fatty acids) and absolute concentrations (mg/g) to summarize dietary and infection effects in Atlantic salmon fry. FA composition data were divided into non-infected, SAV-WT-infected (20 DPC), and SAV-Att-infected (28 DPC) groups, each fed diets with varying EPA/DHA ratios (0.2 E/D, 0.9 E/D, and 2.2 E/D). Key FA patterns are shown in Figures 2, 3 for the SAV-WT and SAV-Att groups, respectively, while full FA profiles (%) and absolute concentrations (mg/g) are provided in Supplementary Tables 4–8.

Figure 2

Figure 3

3.3.1 FA composition of non-infected fish

In this trial, salmon fry received the experimental diets immediately after yolk-sac absorption. The initial FA composition of the fry (% of TFA; Supplementary Table 4; Figures 2, 3) therefore reflects the maternal FA profile derived from the broodstock. After the initiation of feeding, the whole-body FA composition largely reflected the dietary profiles. The pre-feeding resulted in reduced EPA for fish fed 0.2 E/D and 0.9 E/D, but increased EPA in fish fed 2.2 E/D, compared to the initial maternally derived yolk-sac FA composition (Figures 2, 3). DHA increased only in fish fed 0.2 E/D, while it decreased for the other diet groups. The sum of EPA and DHA decreased for all diet groups. ARA decreased for all diet groups when comparing the 0-DPC samples to the yolk-sac fry (Supplementary Table 4; Figures 2, 3). The main shift in FA profile occurred during the pre-feeding period of 28 days, after which the profiles were more stable, with some exceptions. DHA levels increased in fish fed 0.9 E/D and 0.2 E/D from 20 DPC to 28 DPC, while fish fed 2.2 E/D demonstrated continuously reducing DHA levels throughout the study (Figures 2, 3).

EPA closely mirrored dietary levels but was consistently lower in fish tissue than in the diet, reaching approximately half of dietary proportions across all groups (Table 2; Figures 2, 3). Fish fed 2.2 E/D were the only fish to maintain higher EPA levels than in the initial yolk-sac profile (Figures 2, 3). DHA also reflected dietary levels, but retention differed: fish fed 0.2 E/D had tissue DHA nearly identical to the diet, whereas those on 0.9 E/D and 2.2 E/D had tissue levels ~1.2- and ~1.5-fold higher DHA than the diet, respectively (Figures 2, 3). DPA increased with a higher dietary EPA/DHA ratio, despite the dietary DPA decreasing as the ratio increased (Figures 2, 3). ARA levels nearly doubled compared to dietary levels in fish fed 0.2 E/D, while remaining close to dietary levels in fish fed 0.9 E/D and 2.2 E/D.

3.3.2 FA composition of SAV-WT-infected fish

The most striking effect of SAV-WT on FA was the reduced total FA content (mg/g) (Supplementary Table 5; Figure 2), which was particularly prominent in fish fed 0.2 E/D and 2.2 E/D. There were reductions in the absolute levels of almost every single FA, but none were significant for 0.9 E/D—only in the other two groups; 18:1n−9 was the most “expendable”, with the largest losses (p = 0.048 for 0.2 E/D, ns for 0.9 E/D, and p = 0.003 for 2.2 E/D). There were numerical losses of EPA following infection for all groups, but it was only significant in fish fed 2.2 E/D (interaction p < 0.001). DPA was reduced for infected fish fed 0.2 E/D and 2.2 E/D (interaction p = 0.013) compared to controls, and similarly, there were lower levels of DHA in 0.2 E/D and 2.2 E/D (p = 0.064 and p = 0.074, respectively). The main disease effects are visualized in Figure 2. There was a significant interaction effect also for ARA (p = 0.002), with significant reductions after infection seen only for fish fed 2.2 E/D.

In terms of FA profile (% of TFA), the data showed a slightly different trend, as DHA increased in infected fish (Supplementary Table 6; Figure 2). Meanwhile, the share of EPA increased in the 0.2 E/D and 0.9 E/D diet groups but decreased for fish given 2.2 E/D (interaction p = 0.004). Most MUFAs decreased slightly, resulting in a reduced total MUFA. Otherwise, the sum SFA and sum n−3 increased significantly with disease, whereas sum n−6 was unaffected.

3.3.3 FA composition of SAV-Att-infected fish

Contrary to SAV-WT-infected fish, there were hardly any effects of SAV-Att infection on the FA profile. Total FA levels (mg/g) were similar before and after infection, and only a few minor MUFAs differed with infection (Supplementary Table 7; Figure 3). Notably, infection did not cause any distinct drops in the concentrations of EPA, DPA, DHA, or ARA. When looking at the FA profile in % of TFA, there were a few individual MUFAs significantly affected by infection (Supplementary Table 8), but no effects on EPA, DPA, or DHA (Figure 3). There was a small but significant disease-related increase in ARA in 2.2 E/D (Figure 3), which arose from changes in other FA’s rather than numerical changes in absolute ARA content.

3.4 Viral load

The viral load decreased significantly over time for both the SAV-WT (7 DPC > 14 DPC > 20 DPC; Figure 4A) and SAV-Att (7 DPC > 14 DPC > 28 DPC; Figure 4B) groups. Viral load was high with low inter-individual variation already at 7 DPC for SAV-WT, indicating rapid establishment of viral infection in all fish. In contrast to the SAV-WT-infected fish, the SAV-Att group exhibited a significantly lower and more variable viral load among individual fish at 7 DPC, but with the variation diminishing over time. Additionally, there was a significant three-way interaction (p = 0.024), with a significantly higher viral load in fish fed 0.2 E/D compared to 0.9 E/D and 2.2 E/D at 7 DPC only in fish challenged with SAV-Att (Figure 4B).

Figure 4

3.5 Histopathology

From 14 to 21 DPC, histopathology scores increased significantly in the heart, decreased significantly in the pancreas, and showed a non-significant decrease in the muscle. The average scores were significantly different between organs, with the highest scores in the heart, intermediate in the pancreas, and the lowest in the muscle. For SAV-WT, there were no statistically significant effects of diet on heart histopathology (Figure 5A). However, pancreatic histopathology scores in fish fed diets 0.2 E/D and 2.2 E/D tended to decrease from 14 DPC to 20 DPC (Figure 3B, p = 0.0825 and p = 0.0931 for 0.2 E/D and 2.2 E/D, respectively), whereas they remained high in fish fed the 0.9 E/D diet. A similar pattern was observed in muscle tissue, where fish fed the 0.9 E/D diet showed an average lesion score of approximately 2, compared to an average of 1 for fish fed the other two diets at 20 DPC (Figure 5C). No significant differences were found for histopathology in SAV-Att-infected fish (Figures 5D–F). Histopathology scores were overall lower in SAV-Att-infected fish compared to SAV-WT (Figure 5).

Figure 5

4 Discussion

Most studies performed to determine EPA and DHA in aquaculture fish to date have examined the combined effects of EPA and DHA intake in Atlantic salmon. This is primarily because FO, the traditional source of these FA, contains them in relatively similar proportions (Turchini et al., 2009). However, should FO be fully or partially replaced with alternative EPA and DHA sources, the dietary EPA/DHA ratio could vary considerably because many of these novel sources have imbalanced contents of EPA and DHA (Sprague et al., 2016). Since such dietary changes have been reported to affect the immunocompetence of salmon (Martinez-Rubio et al., 2012; Holen et al., 2018; Lutfi et al., 2023; Caballero-Solares et al., 2024), the present study investigated whether manipulating the dietary EPA/DHA ratio, while maintaining constant total EPA+DHA levels, could influence host responses to SAV infection, including viral load dynamics and tissue-level inflammatory processes. Also, the current study demonstrates that dietary EPA/DHA ratios significantly modulate aspects of disease progression, including viral kinetics and lipid metabolism, highlighting the role of FA balance in modulating fish robustness. Furthermore, this study validates the salmon fry disease model as a sensitive and practical tool for evaluating dietary impacts on infectious disease resistance.

4.1 Dietary effects on growth

A recent review by Glencross et al. (2025) reported that no optimal dietary EPA/DHA ratio for growth has been established for salmonids. In most of the studies included in that review, the total EPA+DHA varied among treatments, thereby confounding the specific effect of the ratio itself. To the best of our knowledge, Santigosa et al. (2023) was the only cited study that maintained a constant total EPA+DHA level while varying the EPA/DHA ratio in a trial with salmon. However, it tested only ratios below 1 (i.e., DHA-dominant diets) and found no significant effects on growth. In the present trial, by maintaining a constant total EPA+DHA content while also including EPA/DHA ratios > 1, we address key knowledge gaps identified in the existing literature. Growth performance of salmon fry in the present study decreased with increasing dietary EPA/DHA ratios after the pre-feeding period, which is consistent with results obtained in salmon post-smolts (Hundal et al., 2026). As this effect of dietary EPA/DHA ratio was observed in both fry and post-smolts, it suggests that the effects of ratio on growth are consistent across life stages.

There are several possible mechanistic effects that may have contributed to the observed reduced growth in the high-EPA/low-DHA diets. Earlier trials report a high activation of FA elongation and desaturation pathways in salmonids fed diets with a skewed ratio of EPA/DHA (Betancor et al., 2015; Hong et al., 2022; Hundal et al., 2026). In the present study, the whole-body DHA levels for fish fed 2.2 E/D were 1.6 times higher than in their respective diet, suggesting a compensatory biosynthesis of DHA. Conversely, fry fed the 0.2 E/D diet had body levels of DHA that were comparable to the diet level, indicating adequate dietary DHA supply. Further evidence of active elongation and desaturation in fish given less DHA compared to EPA was seen in the tissue levels of the intermediary FA DPA, which increased with dietary EPA levels despite the overall low DPA content in the diets. An increased metabolic cost of this activation could have caused the reduced growth in the 2.2 E/D diet in the current study, as also indicated to have caused reduced growth in the study by Hundal et al. (2026).

It is also possible that the diets evoked changes in appetite, leading to reduced feed intake in the groups with lower growth. As we could not accurately measure feed intake in the current experimental setup, we cannot exclude these effects, but earlier trials on bigger fish have not reported an effect of skewed EPA/DHA ratios on feed intake (Santigosa et al., 2023; Hundal et al., 2026). Another possibility is that the high EPA content in the 2.2 E/D diet increased FA diversion to β-oxidation rather than storage, as EPA is more readily β-oxidized than DHA (Norambuena et al., 2015; Emery et al., 2016). The lower tissue contents of EPA compared to dietary contents in the 2.2 E/D groups could support this, but it is not possible to determine this with high confidence due to the lack of feed intake data. There is also a report on low dietary DHA contents leading to increased intestinal inflammation in salmonids (Bou et al., 2017), which could reduce the intestinal efficiency for uptake of nutrients. An improved intestinal health was also reported by Løvmo et al. (2022) after feeding a high-EPA diet. This parameter was, however, not in the scope of this study but could be an interesting avenue for future studies.

Taken together, the present results challenge the prevailing view that the dietary EPA/DHA ratio has no effect on growth. While none of the diets in our trial were designed to be critically low in DHA (lowest at 9% of TFA in 2.2 E/D), the growth impairment in this group suggests that dietary DHA supply was nonetheless inadequate for salmon fry. Seeing as most other FA parameters were kept highly consistent between diets, these findings provide a stronger basis for attributing the effect to the dietary EPA/DHA ratio, or perhaps insufficient DHA supply. Overall, the results indicate that optimizing the dietary EPA/DHA ratio is critical for supporting growth in salmon fry.

4.2 Comparison of the two viral strains for nutrition studies

The present study demonstrated distinct differences in virulence between the two viral strains. The SAV-WT was highly virulent, causing rapid and significant mortality shortly before 20 DPC, necessitating premature trial termination. Pronounced histopathological lesions in the heart and pancreas confirmed advanced stages of disease. In contrast, infection with SAV-Att caused no mortality and substantially milder histopathological changes. The whole-body FA composition of SAV-WT-infected fry showed striking changes, whereas SAV-Att infection was accompanied only by minimal shifts in FA composition. This is as expected, as the SAV-WT severely affects the pancreas and thus impacts feed intake and especially lipid digestibility due to the crucial role of the pancreas in lipid digestion. The SAV-Att did not inflict the same damage to the pancreas. Despite the hypothesis that SAV-WT virulence would mask dietary effects, differences between dietary groups were observed in both histopathology and FA content. Diet-induced differences in viral load were, however, only detected in SAV-Att-challenged fish. Although the viral load in the SAV-Att groups was lower than in the SAV-WT group, the infection was nonetheless considered successful for the SAV-Att group, as none of the >100 fish sampled from this group at 14 DPC tested negative.

Both viral strains, therefore, represent useful experimental models. SAV-WT reflects high-virulence infections encountered in aquaculture, making it relevant to industry conditions. However, its rapid disease progression and associated mortality limit its utility for mechanistic dietary studies. The SAV-Att strain provides a milder, non-lethal infection model, enabling the study of both innate and adaptive immune responses, and the resolution of inflammation and disease. This model is particularly useful for investigating the sometimes subtle, longer-term effects of dietary components.

4.3 Effect of disease on tissue fatty acid profile

Infection with SAV-WT induced striking reductions in TFA in the fish tissues compared to the non-infected fish, clearly demonstrating disease-related metabolic disruption, consistent with observations from other salmon disease studies (Herath et al., 2016; Røsæg et al., 2019; Ivanova et al., 2024). This is likely linked to reduced feed intake, a common symptom of pancreatic damage during SAV infections (Taksdal et al., 2007). Feed intake was not quantified in the current study due to the experimental facility setup, so a follow-up trial with a setup adapted for the collection of uneaten feed would have to be performed to confirm whether the changes in TFA seen in the SAV-WT-infected fish were indeed due to reduced feed intake. There were, however, also some interesting differences between the dietary groups. Absolute concentrations of most FAs decreased after infection, with the largest losses in fish fed 0.2 E/D and 2.2 E/D and only minor reductions in fish fed 0.9 E/D. The FAs with the largest reductions in absolute amounts were MUFAs, such as oleic acid, while the lowest reductions were seen in the PUFAs such as DHA and ARA. The selective retention of some FAs, supported by the relative increases in DHA and ARA, suggests the prioritization of these FAs during the SAV-WT infection. This aligns with reports that ARA and DHA are preferentially incorporated into fish membranes and mobilized for lipid inflammatory mediator production rather than oxidation during infection (Mendivil, 2021; Ferreira et al., 2022). This prioritization of long-chained PUFA is also seen in other teleosts during starvation (Arslan et al., 2021), so it is not clear whether this is an effect of stunted feed intake or of the infection itself. The distinctive EPA decline in the 2.2 E/D diet group, both absolute and relative, during the SAV-WT infection is noteworthy. This group was provided the feed with the highest EPA but lowest DHA and further showed the greatest relative DHA increase after SAV-WT infection, which indicates additional DHA synthesis during disease for this group. Collectively, the changes to FA profile indicate that the dietary EPA/DHA ratio influenced how FAs were utilized during a viral infection.

In contrast to SAV-WT, infection with SAV-Att induced only minor changes in absolute FA composition, including TFA, with patterns largely reflecting dietary influences rather than a strong infection-driven metabolic response. This indicates a weak activation of lipid reserves, resembling an attenuated version of the extreme lipid loss observed for the SAV-WT-infected fish. The limited changes in FA profile align with the reduced virulence of SAV-Att and reflect the degree of pancreas damage expected in SAV-Att compared to WT-SAV, as also confirmed by the reduced viral load and lesion severity found. Although FA profiles of the SAV-Att-infected fish showed limited change with infection, diet-driven differences could still have modulated lipid mediator production during disease (Mendivil, 2021; Ferreira et al., 2022). Future studies should therefore quantify EPA- and DHA-derived metabolites to clarify any dietary impacts on a low-grade infection as this model studies.

4.4 Dietary FA profiles and their effects on disease progression

Viral loads in the salmon fry revealed interactions between viral virulence and dietary treatment. For both strains, viral loads declined significantly over time, reflecting effective viral clearance or decay after IP injection. The pattern in viral loads post-infection may look different if using other methods for administering the virus to the fish, such as bath and cohabitation challenge (Kannimuthu et al., 2026). SAV-WT-challenged fish uniformly had high viral loads across diet groups at each time point, whereas SAV-Att challenge resulted in lower viral load means with greater inter-individual variability, especially in the early phase of the infection (i.e., 7 DPC). This difference may reflect the distinct virulence profiles of the two virus strains, where the highly virulent SAV-WT drives uniform infection kinetics across individuals, whereas the less virulent SAV-Att permits individual host factors, such as genetics or fitness, to initially influence infection progression. This difference aligns with reports of reduced virulence in other attenuated SAV strains (Braaen et al., 2025). The SAV-Att fish fed the 0.2 E/D diet displayed significantly higher viral loads than the other groups at 7 DPC, although this did not translate into higher mortality or histopathology scores as the disease progressed. Fish fed diet 2.2 E/D and infected with SAV-WT had the highest (non-significant) mortality, but without clear indications of corresponding viral loads or histopathology scores. Interestingly, viral loads in all SAV-WT-challenged diet groups declined significantly from 7 to 14 DPC, i.e., prior to the onset of mortality, and this trend persisted at 20 DPC, when mortality was likely near its peak. This suggests that viral load magnitude does not correlate directly with mortality magnitude, but rather that peaks in viral load probably precede the onset and peak of mortality rates. In contrast, histopathology scores increased across the SAV-WT-challenged diet groups between 14 and 20 DPC, suggesting a better temporal correlation with observed mortality. The lack of a clear connection between viral load and disease progression is also shown in earlier studies, both in vivo (Martinez-Rubio et al., 2012; Hatlen et al., 2016) and in vitro (Holmlund et al., 2025). However, as cumulative mortality was not allowed to progress undisturbed due to sampling constraints, our data likely do not reflect the final mortality rates. Instead, the observed mortalities may indicate differences in the onset and kinetics of mortality between the diet groups. These observations highlight the need for future work to disentangle the mechanisms through which diet and host factors modulate pathology independently of viral replication dynamics.

The dietary responses seen in the present study varied through the disease progression, with differing effects of dietary FA composition on the initial phases of the disease compared to the later phases. The 0.2 E/D group, having overall the highest viral loads, especially in the early phase, of both the SAV-WT and the SAV-Att infections, also showed signs of improved recovery in the SAV-WT-challenged fish compared to fish fed the diet with a more balanced ratio (0.9 E/D). While DHA supports resolution through its role as a precursor of several anti-inflammatory SPMs (Zuo et al., 2012; Xu et al., 2016; Bou et al., 2020; Ferreira et al., 2022), insufficient EPA has been reported to weaken early innate immune signaling (Peterson et al., 1998; Zuo et al., 2012; Xu et al., 2016; Gjøen et al., 2024). This could explain why the EPA-poor/DHA-rich 0.2 E/D diet led to initially high viral loads (due to low EPA) but an improved disease resolution (due to high DHA).

The other diet with a skewed EPA/DHA ratio, 2.2 E/D, was the group with the highest registered mortality after the SAV-WT infection. This dietary group was also reported to have similar positive effects on disease in the later stages of the trial as observed in the 0.2 E/D group, but without the generally higher viral loads. This could be due to this diet having a positive effect on disease resolution, or merely that the sickest (deceased) fish were not part of the pool of fish available for sampling, thus giving an unrepresentative population at the last sampling point. If this slight improvement in histopathology is a true dietary effect, this is consistent with the recovery benefits reported for high-EPA diets in Atlantic salmon (Hatlen et al., 2016) and with the enhanced survival of ISAV-infected cells under EPA stimulation (Holmlund et al., 2025). The results on viral load are, however, in contradiction with those of Hatlen et al. (2016), who reported a higher initial viral load in salmon fed a high EPA diet. It is therefore possible that the high-EPA diet led to an improved recovery, but due to the high mortality in this group, this is more difficult to conclude based on the current study.

The 0.9 E/D diet showed poorer recovery than the diets with a more skewed EPA/DHA ratio. This group also had only minor changes in tissue FA composition after infection with SAV-WT, contrasting the marked FA losses seen in fish fed 0.2 E/D and 2.2 E/D. It is possibly reflecting that the other dietary groups better mobilized FAs for energetic requirements during disease; however, this remains unclear.

The relationship between EPA/DHA ratio and viral load appears variable across studies and contexts (Martinez-Rubio et al., 2012; Hatlen et al., 2016; Holmlund et al., 2025), requiring further investigations to resolve the mechanisms. Nonetheless, the current findings indicate that dietary FA ratios shape both viral kinetics and tissue-level responses through distinct mechanisms, underlining the importance of considering EPA- and DHA-derived functions separately when studying FA impacts on viral diseases in salmonids.

5 Conclusions

Increasing the dietary EPA/DHA ratio while maintaining constant total EPA+DHA significantly reduced growth in non-infected Atlantic salmon fry, indicating that EPA and DHA themselves, whether through an imbalanced ratio or insufficient DHA, are the likely drivers of this effect. Effects of varying the EPA/DHA ratio were also evident following infection with both the wild-type and attenuated SAV strains, demonstrating that the dietary EPA/DHA ratio influences growth performance and modulates specific host responses to viral infection, including viral load dynamics and lipid metabolism. However, the dietary effects on viral load, tissue inflammation, and survival were distinct and sometimes opposing between the different diets, making it difficult to clearly conclude on which diet was optimal for the salmon fry in this specific scenario.

As experimental models for nutrition studies, SAV-WT produced severe disease with high mortality due to tissue damage and large reductions in lipid storage, limiting the study duration but still revealing interesting dietary effects. SAV-Att induced only mild disease, with less tissue damage and fewer effects on lipid storage, where dietary effects on viral load dynamics were evident but not detected in histology or FA profile.

Longer follow-up studies may be needed to capture dietary effects on the recovery phase for SAV-Att. Future studies should separate dietary effects on viral replication from those on recovery. In addition, the analyses of lipid-derived inflammatory mediators could reveal disease-related effects of altered FA profiles not evident from bulk composition but still driven by it.

References

• 1

  AasT. S.ÅsgårdT.YtrestøylT. (2022). Utilization of feed resources in the production of rainbow trout (Oncorhynchus mykiss) in Norway in 2020. Aquacult. Rep.26, 101317. doi: 10.1016/j.aqrep.2022.101317

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 2

  AksnesI.BraaenS.MarkussenT.ÅkessonC. P.VilloingS.RimstadE. (2021). Genetically modified attenuated salmonid alphavirus: A potential strategy for immunization of Atlantic salmon. J. Fish Dis.44, 923–937. doi: 10.1111/jfd.13352

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 3

  AraujoP.EspeM.HolenE.AustgulenM. H.IqbalS.HundalB. K. (2025). Cross-species comparison of resolvin E and resolvin D biosynthesis: Quantification by liquid chromatography mass spectrometry in fish and human cells exposed to alpha-linolenic acid, eicosapentaenoic acid, and docosahexaenoic acid. J. Chromatogr. B1264, 124729. doi: 10.1016/j.jchromb.2025.124729

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 4

  AraujoP.LucenaE.YangY.CeemalaB.MengeshaZ.HolenE. (2014). The impact of exogenous ω-6 and ω-3 polyunsaturated fatty acids on the induced production of pro- and anti-inflammatory prostaglandins and leukotrienes in Atlantic salmon head kidney cells using a full factorial design and LC–MS/MS. J. Chromatogr. B964, 164–171. doi: 10.1016/j.jchromb.2014.01.018

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 5

  ArslanG.BayırM.YağanoğluA. M.BayırA. (2021). Changes in fatty acids, blood biochemistry and mRNA expressions of genes involved in polyunsaturated fatty acid metabolism in brown trout (Salmo trutta) during starvation and refeeding. Aquacult. Res.52, 494–504. doi: 10.1111/are.14908

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 6

  BasilM. C.LevyB. D. (2016). Specialized pro-resolving mediators: endogenous regulators of infection and inflammation. Nat. Rev. Immunol.16, 51–67. doi: 10.1038/nri.2015.4

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 7

  BatesD.MächlerM.BolkerB.WalkerS. (2015). Fitting linear mixed-effects models using lme4. J. Stat. Software67, 1–48. doi: 10.18637/jss.v067.i01

  • CrossRef
  • Google Scholar

• 8

  BetancorM. B.SpragueM.SayanovaO.UsherS.CampbellP. J.NapierJ. A.et al. (2015). Evaluation of a high-EPA oil from transgenic Camelina sativa in feeds for Atlantic salmon (Salmo salar L.): Effects on tissue fatty acid composition, histology and gene expression. Aquaculture444, 1–12. doi: 10.1016/j.aquaculture.2015.03.020

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 9

  BetancorM. B.SpragueM.SayanovaO.UsherS.MetochisC.CampbellP. J.et al. (2016). Nutritional evaluation of an EPA-DHA oil from transgenic Camelina sativa in feeds for post-smolt Atlantic salmon (Salmo salar L.). PloS One11, e0159934. doi: 10.1371/journal.pone.0159934

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 10

  BouM.BergeG. M.BaeverfjordG.SigholtT.ØstbyeT.-K.RomarheimO. H.et al. (2017). Requirements of n-3 very long-chain PUFA in Atlantic salmon (Salmo salar L): Effects of different dietary levels of EPA and DHA on fish performance and tissue composition and integrity. Br. J. Nutr.117, 30–47. doi: 10.1017/s0007114516004396

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 11

  BouM.TorgersenJ. S.ØstbyeT.-K. K.RuyterB.WangX.ŠkugorS.et al. (2020). DHA modulates immune response and mitochondrial function of Atlantic salmon adipocytes after LPS treatment. Int. J. Mol. Sci.21, 4101. doi: 10.3390/ijms21114101

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 12

  BraaenS.WesselØ.BjørgenH.RimstadE. (2025). Long-term protection in Atlantic salmon (Salmo salar) to pancreas disease (PD) can be achieved through immunization with genetically modified, live attenuated salmonid alphavirus 3. Vaccines13, 1–17. doi: 10.3390/vaccines13020190

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 13

  Caballero-SolaresA.EslamlooK.HallJ. R.KatanT.EmamM.XueX.et al. (2024). Vegetable omega-3 and omega-6 fatty acids differentially modulate the antiviral and antibacterial immune responses of Atlantic salmon. Sci. Rep.14, 10947. doi: 10.1038/s41598-024-61144-w

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 14

  Caballero-SolaresA.HallJ. R.XueX.EslamlooK.TaylorR. G.ParrishC. C.et al. (2017). The dietary replacement of marine ingredients by terrestrial animal and plant alternatives modulates the antiviral immune response of Atlantic salmon (Salmo salar). Fish. Shellfish Immunol.64, 24–38. doi: 10.1016/j.fsi.2017.02.040

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 15

  CalderP. C. (2017). Omega-3 fatty acids and inflammatory processes: from molecules to man. Biochem. Soc Trans.45, 1105–1115. doi: 10.1042/bst20160474

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 16

  ChristensenR. (2025). ordinal-Regression Models for Ordinal Data. R package version 2025.12-29.

  • Google Scholar

• 17

  EmeryJ. A.NorambuenaF.TrushenskiJ.TurchiniG. M. (2016). Uncoupling EPA and DHA in fish nutrition: Dietary demand is limited in Atlantic salmon and effectively met by DHA alone. Lipids51, 399–412. doi: 10.1007/s11745-016-4136-y

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 18

  FerreiraI.FalcatoF.BandarraN.RauterA. P. (2022). Resolvins, protectins, and maresins: DHA-derived specialized pro-resolving mediators, biosynthetic pathways, synthetic approaches, and their role in inflammation. Molecules27, 1677. doi: 10.3390/molecules27051677

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 19

  FoxJ.WeisbergS. (2019). An R Companion to Applied Regression (Thousand Oaks, California: Sage Publications).

  • Google Scholar

• 20

  FringuelliE.RowleyH. M.WilsonJ. C.HunterR.RodgerH.GrahamD. A. (2008). Phylogenetic analyses and molecular epidemiology of European salmonid alphaviruses (SAV) based on partial E2 and nsP3 gene nucleotide sequences. J. Fish Dis.31, 811–823. doi: 10.1111/j.1365-2761.2008.00944.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 21

  GjøenT.RuyterB.ØstbyeT. K. (2024). Effects of eicosapentaneoic acid on innate immune responses in an Atlantic salmon kidney cell line in vitro. PloS One19, e0302286. doi: 10.1371/journal.pone.0302286

  • CrossRef
  • Google Scholar

• 22

  GlencrossB. D.BachisE.BetancorM. B.CalderP.LilandN.NewtonR.et al. (2025). Omega-3 futures in aquaculture: Exploring the supply and demands for long-chain omega-3 essential fatty acids by aquaculture species. Rev. Fisheries Sci. Aquaculture33, 167–216. doi: 10.1080/23308249.2024.2388563

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 23

  GlencrossB. D.EnricoB.B.M.PhilipC.LilandN.NewtonR.et al. (2024). Omega-3 futures in aquaculture: Exploring the supply and demands for long-chain omega-3 essential fatty acids by aquaculture species. Rev. Fisheries Sci. Aquaculture33, 167–216. doi: 10.1080/23308249.2024.2388563

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 24

  GlencrossB. D.TocherD. R.MatthewC.BellJ. G. (2014). Interactions between dietary docosahexaenoic acid and other long-chain polyunsaturated fatty acids on performance and fatty acid retention in post-smolt Atlantic salmon (Salmo salar). Fish Physiol. Biochem.40, 1213–1227. doi: 10.1007/s10695-014-9917-8

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 25

  HarrellF. E. (2015). “ Assumptions and interpretation of parameters. Ordinal Logistic Regression,” in Regression Modeling Strategies ( Springer, Cham).

  • Google Scholar

• 26

  HatlenB.JørgensenS. M.TimmerhausG.KrasnovA.BouM.RuyterB.et al. (2016). “ Styrt fettsyresammensetning i fôr for å forebygge utbrudd av viktige virussykdommer i laks,” ( Norwegian University of Life Sciences (NMBU, Ås, Norway). Nofima report 3/2016; pp 1-47. Available online at: https://nva.sikt.no/registration/0198cc5c8e08-54f41480-5b5f-4311-907e-99ca738dc035.

  • Google Scholar

• 27

  HerathT. K.FergusonH. W.WeidmannM. W.BronJ. E.ThompsonK. D.AdamsA.et al. (2016). Pathogenesis of experimental salmonid alphavirus infection in vivo: an ultrastructural insight. Vet. Res.47, 7. doi: 10.1186/s13567-015-0300-2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 28

  HodnelandK.EndresenC. (2006). Sensitive and specific detection of Salmonid alphavirus using real-time PCR (TaqMan). J. Virol. Methods131, 184–192. doi: 10.1016/j.jviromet.2005.08.012

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 29

  HolenE.AraujoP.SissenerN. H.RosenlundG.WaagbøR. (2018). A comparative study: Difference in omega-6/omega-3 balance and saturated fat in diets for Atlantic salmon (Salmo salar) affect immune-, fat metabolism-, oxidative and apoptotic-gene expression, and eicosanoid secretion in head kidney leukocytes. Fish. Shellfish Immunol.72, 57–68. doi: 10.1016/j.fsi.2017.10.040

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 30

  HolenE.HeJ. Y.EspeM.ChenL. Q.AraujoP. (2015). Combining eicosapentaenoic acid, decosahexaenoic acid and arachidonic acid, using a fully crossed design, affect gene expression and eicosanoid secretion in salmon head kidney cells in vitro. Fish. Shellfish Immunol.45, 695–703. doi: 10.1016/j.fsi.2015.05.017

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 31

  HolmlundI.AhmadiS.RuyterB.ØstbyeT.-K.BouM.GjøenT. (2025). Effect of eicosapentaenoic acid on innate immune responses in Atlantic salmon cells infected with infectious salmon anemia virus. Virol. J.22, 5. doi: 10.1186/s12985-024-02619-0

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 32

  HongJ.BledsoeJ. W.OverturfK. E.LeeS.IassonovaD.SmallB. C. (2022). LatitudeTM oil as a sustainable alternative to dietary fish oil in rainbow trout (Oncorhynchus mykiss): Effects on filet fatty acid profiles, intestinal histology, and plasma biochemistry. Front. Sustain. Food Syst.6. doi: 10.3389/fsufs.2022.837628

  • CrossRef
  • Google Scholar

• 33

  HundalB. K.LilandN. S.RosenlundG.HöglundE.AraujoP.StubhaugI.et al. (2021). Increasing the dietary n-6/n-3 ratio alters the hepatic eicosanoid production after acute stress in Atlantic salmon (Salmo salar). Aquaculture534, 736272. doi: 10.1016/j.aquaculture.2020.736272

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 34

  HundalB. K.SissenerN. H.VigenJ.PinoJ.LilandN. S. (2026). EPA alone cannot sustain the requirement for EPA + DHA in Atlantic salmon (Salmo Salar L.): Insights into the role of EPA:DHA ratio and varying EPA levels on growth and tissue fatty acid composition. Aquaculture613, 743449. doi: 10.1016/j.aquaculture.2025.743449

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 35

  IvanovaL.Rangel-HuertaO. D.TartorH.DahleM. K.UhligS.FæsteC. K. (2024). Metabolomics and multi-omics determination of potential plasma biomarkers in PRV-1-infected Atlantic salmon. Metabolites14, 1–21. doi: 10.3390/metabo14070375

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 36

  JorgeD. M. M.MillsR. E.LauringA. S. (2015). CodonShuffle: a tool for generating and analyzing synonymously mutated sequences. Virus Evol.1, vev012. doi: 10.1093/ve/vev012

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 37

  KannimuthuD.RohH.MortonH. C.VossgårdA.PeñarandaM. M. D.FjelldalP. G.et al. (2026). The susceptibility of Atlantic salmon and brown trout to salmonid alphavirus subtype 3 across fry, parr and post-smolt stages. Aquaculture613, 743300. doi: 10.1016/j.aquaculture.2025.743300

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 38

  KuznetsovaA.BrockhoffP. B.ChristensenR. H. B. (2017). lmerTest package: Tests in linear mixed effects models. J. Stat. Software82, 1–26. doi: 10.18637/jss.v082.i13

  • CrossRef
  • Google Scholar

• 39

  LenthR. (2025). emmeans: estimated marginal means, aka least-squares means, comprehensive R archive network (CRAN). doi: 10.32614/cran.package.emmeans

  • CrossRef
  • Google Scholar

• 40

  LiuM.ZhouX.LiY.MaS.PanL.ZhangX.et al. (2022). TIGAR alleviates oxidative stress in brain with extended ischemia via a pentose phosphate pathway-independent manner. Redox Biol.53, 102323. doi: 10.1016/j.redox.2022.102323

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 41

  LøvmoS. D.SundhH.WhatmoreP.NordviM. F.SigholtT.MadaroA.et al. (2022). Intestinal health in Atlantic salmon post-smolt (Salmo salar) when fed low- and high HUFA diets. Aquaculture557, 738318. doi: 10.1016/j.aquaculture.2022.738318

  • CrossRef
  • Google Scholar

• 42

  LutfiE.BergeG. M.BæverfjordG.SigholtT.BouM.LarssonT.et al. (2023). Increasing dietary levels of the n-3 long-chain PUFA, EPA and DHA, improves the growth, welfare, robustness and fillet quality of Atlantic salmon in sea cages. Br. J. Nutr.129, 10–28. doi: 10.1017/s0007114522000642

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 43

  MaJ.XuC.ZhouY.JiangN.XueM.CaoJ.et al. (2023). Metabolomics in rare minnow (Gobiocypris rarus) after infection by attenuated and virulent grass carp reovirus genotype II. Fish. Shellfish Immunol.138, 108840. doi: 10.1016/j.fsi.2023.108840

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 44

  Martinez-RubioL.EvensenØ.KrasnovA.JørgensenS. M.WadsworthS.RuohonenK.et al. (2014). Effects of functional feeds on the lipid composition, transcriptomic responses and pathology in heart of Atlantic salmon (Salmo salar L.) before and after experimental challenge with Piscine Myocarditis Virus (PMCV). BMC Genomics15, 462. doi: 10.1186/1471-2164-15-462

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 45

  Martinez-RubioL.MoraisS.EvensenØ.WadsworthS.RuohonenK.VecinoJ. L. G.et al. (2012). Functional feeds reduce heart inflammation and pathology in Atlantic salmon (Salmo salar L.) following experimental challenge with Atlantic salmon reovirus (ASRV). PloS One7, e40266. doi: 10.1371/journal.pone.0040266

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 46

  McLoughlinM. F.GrahamD. A.NorrisA.MatthewsD.FoyleL.RowleyH. M.et al. (2006). Virological, serological and histopathological evaluation of fish strain susceptibility to experimental infection with salmonid alphavirus. Dis. Aquat. Org.72, 125–133. doi: 10.3354/dao072125

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 47

  MendivilC. O. (2021). Dietary fish, fish nutrients, and immune function: A review. Front. Nutr.7. doi: 10.3389/fnut.2020.617652

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 48

  NorambuenaF.MoraisS.EmeryJ. A.TurchiniG. M. (2015). Arachidonic acid and eicosapentaenoic acid metabolism in juvenile Atlantic salmon as affected by water temperature. PloS One10, e0143622. doi: 10.1371/journal.pone.0143622

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 49

  Norwegian Veterinary Institute (2023). Fiskehelserapporten 2023 (Ås, Norway: Norwegian Veterinary Institute).

  • Google Scholar

• 50

  ObohA.KabeyaN.Carmona-AntoñanzasG.CastroL. F. C.DickJ. R.TocherD. R.et al. (2017). Two alternative pathways for docosahexaenoic acid (DHA, 22:6n-3) biosynthesis are widespread among teleost fish. Sci. Rep.7, 3889. doi: 10.1038/s41598-017-04288-2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 51

  OliverL.DietrichT.MarañónI.VillaránM. C.BarrioR. J. (2020). Producing omega-3 polyunsaturated fatty acids: A review of sustainable sources and future trends for the EPA and DHA market. Resources9, 148. doi: 10.3390/resources9120148

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 52

  OlsvikP.LieK.JordalA.-ENilsenT.HordvikI.. (2005). “ Evaluation of potential reference genes in real-time RT-PCR studies of Atlantic salmon. BMC Molecular Biology6 (1), 21-29. doi: 10.1186/1471-2199-6-21.

  • CrossRef
  • Google Scholar

• 53

  PedersenT. L. (2025). Patchwork: The Composer of Plots. Available online at: https://patchwork.data-imaginist.com/

  • Google Scholar

• 54

  PetersonL. D.JefferyN. M.ThiesF.SandersonP.NewsholmeE. A.CalderP. C. (1998). Eicosapentaenoic and docosahexaenoic acids alter rat spleen leukocyte fatty acid composition and prostaglandin E2 production but have different effects on lymphocyte functions and cell-mediated immunity. Lipids33, 171–180. doi: 10.1007/s11745-998-0193-y

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 55

  PettersonE.StormoenM.EvensenØ.MikalsenA. B.HauglandØ. (2013). Natural infection of Atlantic salmon (Salmo salar L.) with salmonid alphavirus 3 generates numerous viral deletion mutants. J. Gen. Virol.94, 1945–1954. doi: 10.1099/vir.0.052563-0

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 56

  R Core Team (2024). R: A Language and Environment for Statistical Computing ( R Foundation for Statistical Computing). Available online at: https://www.r-project.org/

  • Google Scholar

• 57

  RøsægM. V.RimstadE.GuttvikA.SkjelstadB.BendiksenE.Å.GarsethÅ.H. (2019). Effect of pancreas disease caused by SAV 2 on protein and fat digestion in Atlantic salmon. J. Fish Dis.42, 97–108. doi: 10.1111/jfd.12914

  • CrossRef
  • Google Scholar

• 58

  SantigosaE.OlsenR. E.MadaroA.SøftelandL.CarrI. (2023). The impact of varying EPA:DHA ratio on Atlantic salmon health and welfare. Aquaculture576, 739868. doi: 10.1016/j.aquaculture.2023.739868

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 59

  SerhanC. N.PetasisN. A. (2011). Resolvins and protectins in inflammation resolution. Chem. Rev.111, 5922–5943. doi: 10.1021/cr100396c

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 60

  SissenerN. H.TorstensenB. E.StubhaugI.RosenlundG. (2016). Long-term feeding of Atlantic salmon in seawater with low dietary long-chain n-3 fatty acids affects tissue status of the brain, retina and erythrocytes. Br. J. Nutr.115, 1919–1929. doi: 10.1017/s0007114516000945

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 61

  SpragueM.DickJ. R.TocherD. R. (2016). Impact of sustainable feeds on omega-3 long-chain fatty acid levels in farmed Atlantic salmon 2006–2015. Sci. Rep.6, 21892. doi: 10.1038/srep21892

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 62

  TaksdalT.OlsenA. B.BjerkåsI.HjortaasM. J.DannevigB. H.GrahamD. A.et al. (2007). Pancreas disease in farmed Atlantic salmon, Salmo salar L. and rainbow trout, Oncorhynchus mykiss (Walbaum), in Norway. J. Fish Dis.30, 545–558. doi: 10.1111/j.1365-2761.2007.00845.x

  • CrossRef
  • Google Scholar

• 63

  TocherD. R. (2010). Fatty acid requirements in ontogeny of marine and freshwater fish. Aquacult. Res.41, 717–732. doi: 10.1111/j.1365-2109.2008.02150.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 64

  TocherD. R.BetancorM. B.SpragueM.OlsenR. E.NapierJ. A. (2019). Omega-3 long-chain polyunsaturated fatty acids, EPA and DHA: Bridging the gap between supply and demand. Nutrients11, 1–20. doi: 10.3390/nu11010089

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 65

  TurchiniG. M.TorstensenB. E.NgW.-K. (2009). Fish oil replacement in finfish nutrition. Rev. Aquacult.1, 10–57. doi: 10.1111/j.1753-5131.2008.01001.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 66

  WickhamH. (2016). Ggplot2: Elegant Graphics for Data Analysis (New York: Springer-Verlag New York).

  • Google Scholar

• 67

  WickhamH.AverickM.BryanJ.ChangW.McGowanL. D. A.FrançoisR.et al. (2019). Welcome to the tidyverse. J. Open Source Software4, 1686. doi: 10.21105/joss.01686

  • CrossRef
  • Google Scholar

• 68

  XuH.WangJ.MaiK.XuW.ZhangW.ZhangY.et al. (2016). Dietary docosahexaenoic acid to eicosapentaenoic acid (DHA/EPA) ratio influenced growth performance, immune response, stress resistance and tissue fatty acid composition of juvenile Japanese seabass, Lateolabrax japonicus (Cuvier). Aquacult. Res.47, 741–757. doi: 10.1111/are.12532

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 69

  ZuoR.AiQ.MaiK.XuW.WangJ.XuH.et al. (2012). Effects of dietary docosahexaenoic to eicosapentaenoic acid ratio (DHA/EPA) on growth, nonspecific immunity, expression of some immune related genes and disease resistance of large yellow croaker (Larmichthys crocea) following natural infestation of parasites (Cryptocaryon irritans). Aquaculture334-337, 101–109. doi: 10.1016/j.aquaculture.2011.12.045

  • Pubmed Abstract
  • CrossRef
  • Google Scholar